A sub-bottom profiler (SBP) can capture the sediment interfaces and properties of different types of sediment. Horizon picking from SBP images is one of the most crucial steps in marine sub-bottom sediment interpretation. However, traditional horizon picking methods are good at obtaining the main horizons representing the main reflectors while ignoring the detailed horizons. While detailed horizons are the prime objective, many tiny structures caused by interference echoes will also be picked. To overcome this limitation, an integrated horizon picking method for obtaining the main and detailed horizons simultaneously is proposed in this paper. A total of three main process steps: the diffusion filtering method, the enhancement filtering method as well as the local phase calculation method, are used to help obtain the main and detailed horizons. The diffusion filtering method smooths the SBP images and preserves reflectors. Enhancement filtering can eliminate outliers and enhance reflectors. The local phase can be used to highlight all of the reflections and help in the choosing of detailed horizons. A series of experiments were then performed to validate the effectiveness of the proposed method, and good performances were achieved.

中文翻译：

---

一种用于获取海底剖面仪声纳图像主要和详细反射体的综合地平线拾取方法

海底剖面仪 (SBP) 可以捕获不同类型沉积物的沉积物界面和特性​​。从 SBP 图像中选取地平线是海洋亚底沉积物解释中最关键的步骤之一。然而，传统的层位选取方法善于获取代表主反射体的主要层位，而忽略了详细的层位。虽然详细的地平线是主要目标，但也将拾取由干扰回波引起的许多微小结构。为了克服这一限制，本文提出了一种同时获取主要和详细层位的综合层位拾取方法。共有三个主要处理步骤：扩散滤波法、增强滤波法以及局部相位计算法，用于帮助获得主要和详细的层位。扩散滤波方法平滑 SBP 图像并保留反射器。增强过滤可以消除异常值并增强反射器。局部相位可用于突出所有反射并帮助选择详细的视野。随后进行了一系列实验以验证所提出方法的有效性，并取得了良好的性能。